AlgoPlus//machine learning / rnn
Read the theory

Recurrent Neural Network

One cell reused across time, carrying a hidden state forward as memory.

Timestep
Tokens
Legend
Input token
Cell (shared)
Hidden state
AI Tutor Workspace
In a nutshell
A recurrent neural network reads a sequence one item at a time, keeping a hidden state — a running summary of everything seen so far — and feeding it back in at each step. That loop gives the network memory, so earlier words can shape later predictions. But over long sequences the learning signal fades on its way back (the vanishing gradient problem), so plain RNNs struggle to remember things from far away.
Ready
Press play to begin the cinematic walkthrough.
Read left to right, jotting a running summary you carry to the next word. Same pen every step — but a long sentence smudges the earliest notes.
Key terms
Go deeper in the lesson
Read the full theory, intuition & complexity for Recurrent Neural Network.